RESEARCH DEPARTMENT
N.TeSC, COLOUR TELEVISION SYSTEMS
INTERFERENCE ON MONOCHROME TELEVISION RECEIVERS
BY THE CHROMINANCE SUBCARRIER
Report No. T.05I
Serial No. I 95*/ 37
y^fec^-^^W'
LC. Mann
R o 0=Ao Maurice, Ing.-Dr., A„MohE=E. (Wo Proctor Wilson)
This Report is the property of the
British Broadcasting Corporation and may
not be reproduced or disclosed to a
third party in any form without the
written permission of the Corporation,
Report No. T.051
NVT..S..C COLOUR TELEVISION SYSTEM:
INTERFERENCE ON MONOCHROME TELEVISION RECEIVERS
BY: THE CHROMINANCE SUBCARRIER
Section Title Page
SUMPiRY
INTRODUCTION
METHOD
RESULTS AND CONCLUSIONS
APPENDIX 1
APPENDIX -2
CONFIDENTIAL
December 1954
Report No. T.051
Serial No. 1954/37
N".T.:S..C.: COLOUR TELEVISION" SYSTEM:
INTERFERENCE ON MONOCHROME TELEVISION RECEIVERS
BY: THE CHROMINANCE SUBCARRIER
SUMMARY
The tests described were undertaken during the period in July for which a
simulated chrominance signal was radiated from Alexandra Palace, and, as far as
possible, the radiated signal was used for the purpose of the tests. Measurements
were made at Kingswood on a group of six domestic television receivers, the main
objects being:
a. To measure at the sound output of each receiver the amount of
breakthrough or crosstalk caused by the subcarrier modulation, and
to compare this with the measured background noise which was
audible from the loudspeaker when the receiver was functioning
normally, that is, when no chrominance signal was present*
b. To assess the perceptibility of the dot structure resulting from
subcarrier interference and to compare this with the perceptibility
of the normal raster structure.
The results indicate that in test a the amount of interference on the
sound channels of the receivers was so small as to approximate to the normal break-
through of vision on sound.
In test b the average of the viewers could still perceive the normal
raster lines at a distance twice that at which they could see the dot pattern due to
the subcarrier.
Two subsidiary features have been examined during the preparatory work for
the two main tests which form the substance of this report. They are:
i. A reason for the subcarrier dot pattern appearing brighter than
the mean raster illumination.
ii. The perceptibility of line structure at various brightness levels.
These secondary observations are included as Appendices 1 and 2 respectively.
1. INTRODUCTION.
At the request of the B.R.E.M..A. Colour Television Sub-Committee, a simulated
chrominance signal was radiated out of programme hours from the main Alexandra Palace
vision transmitter between 12 noon and 12,30 p*m. for the period of several days.
It was proposed that this signal should be received on various domestic television
receivers in order to assess the degree of interference that would result with a fully
compatible colour transmissions
Line— up tone was transmitted on the sound channel at the commencement of
each test period so that the receivers could be adjusted to a comfortable listening
level, after which, the sound modulation was removed, leaving carrier only.
IOOt
90--
* 80"
§. 70
E
o 60--
50--
40"
| 30
20"
IO"
o- 1 -
Simuioted chrominance subcarrier
line frequency
2
525 x
~2-65Mc/s
.-a
r^
Simulated luminance
i
n
1
(i
IF
I
-*■ 9-5ps
*» r i-2 utl
A line in upper
half of raster
l
A line in lower
half of raster
Time +•
Fig I - Waveform of test transmission
The waveform of the vision test signal is shown in Pig. 1, and it will he
noticed that the d. a. p. suhcarrier modulation represents approximately 50# of the
total picture and synchronising signal amplitude.
Six domestic receivers of various makes were used for these tests. Three
of these had recently been re-aligned by their manufacturers, whilst the remaining
three were in good working order but had not been re-aligned for some time. One of
the receivers was known to have maladjusted vision and sound rejectors, allowing the
response-curve skirts of both channels to overlap. The R.F.-I.F. sound channel
response curves for three of the receivers used are shown in Fig. 2. The curves
are self-explanatory.
In the subjective test, due to the short period of each transmission, the
observers were limited to seven in number.
Receiver
dB down, at
subcarrier
freq*
dB down
at3Mc/»
off vision
freq.
I
2
3
-42
-32
-43
-21
-16
-34
Average
-39
-23-6
Output measured at sound
detector toad, normalised.
Chrominance
subcarrier
Fig, 2
42 43
Frequency, Mc/s
Sound channel response curves
44
45
46
2* METHOD,
For the objective measurement of subcarrier breakthrough on the sound
channel, it was decided that the input to the test gear should be made by tapping
across the loudspeaker speech-coil of each receiver. This method has some disadvan-
tages which may effect the accuracy of the results to . a slight degree, but the
maintenance of the dynamic loading of the output stages, negative feedback loops and
other considerations, appeared to make any alternative procedure less reliable.
1
Distributee
amplifier
>n
5
I
4
i
1
1
i
1
>
\
r
3
i
L
4
i
i
i
i
J
i
Television receivers
Sound output taps
Aural
Amplifier
Thermal
bridge
/^uA>
netv
fOI
■k'
uneter
Fig* 3 - Apparatus layout for sound tests
The sensitivity and volume controls of each receiver were adjusted to normal
listening level (see Research Department Report No* A*026) and thereafter left at
these settings. At the same time contrast and brightness were adjusted, using Test
Card C. A schematic diagram of the apparatus is shown in Fig. 3; it will be noticed
that the circuit includes a weighting network simulating the response of the human ear.
Initially, the outputs were measured with aerial connected and only sound
carrier present, after which an audio noise measurement was made on each receiver with
an input of picture accompanied by unmodulated sound carrier. Finally, readings were
taken with the chrominance subcarrier superimposed on the vision carrier; the sound
carrier was maintained without modulation. The results are tabulated in Table 1.
TABLE 1
Receiver
Sound Carrier on
(Unmodulated)
No
With
Without
No,
Picture
Picture
Picture,
Signal
Signal
V/ith Subcarrier
dB
dB
dB
1
-52--
-52
-49' 5
2
-45-5
-42
-41
3
-50--5
-44
-46
4
-38-5
*
*
5\
-48-5
-34
-40*5
6
-51
-46-5
-46
Average
of
-47-7
-43-7
-44-6
Receivers
All readings relative to an input of sound carrier modulated with 1000 c/s tone with
receivers adjusted for a sound output of 75 x 10"" 16 watts/ cm \
*Unrelial)le results due to incipient receiver fault.
The second part of the subcarrier tests, that of subjective estimation of
the interference producing a visual dot structure on the screens of the receivers, was
conducted with the experimental transmissions from Alexandra Palace* Reference
should again be made to the transmitted waveform which is shown in Fig. 1.
It will be noticed that the signal used was composed of approximately half a
field of 2*65 Mc/s modulation preceded by approximately half a field of constant grey
level*
Originally, it was expected that this waveform would produce a uniform grey
on the screens of the receivers with dot structure in the lower half. Such, however,
was not the case and a reason for the visual difference is given, as previously
stated, in Appendix 1*
For the purpose of the test, each subject was asked in turn to approach a
receiver and to indicate the position at which he could just perceive the raster
lines, and then to advance or retreat to the point at which he could just discern the
dot structure. The two corresponding distances were measured and recorded.
Due to the difference in brightness between the two halves of the screens,
it was considered at first that some correction to the results might be desirable.
Photometric measurements showed that the dot structure was on an average 40# brighter
than the unmodulated raster.
The perceptibility of the raster structure at various brightness levels was
therefore investigated, and the resultant curve is shown in Appendix 2. Prom this
it can be seen that at screen brightnesses within the range of 5 to 10 ft— lamberts
(54 to 110 asb } the change in distance for the threshold of perception of discrete
raster lines for small changes (such as 40%) of brightness is negligible and falls
within the limits of experimental error for this type of subjective test.
The brightnesses of the receiver screens were therefore adjusted to be, as
far as possible, within this range, and no correction was considered necessary. The
results of the dot— pattern versus raster-^structure visibility tests are shown in
Table 2, as the ratios of distances at the thresholds of perception, of dot structure
to raster structure.
TABLE 2
Observer
Receivers
Average
Assessment
per Observer
for all Receivers
1
2
3
4
5
6
1
0--67-
0-45
0-9
0-38
0-86
0-532
2
0*55
0-55
0- 3
0-5
0-33
0-45
0-446
3
■0-71
0-57
0'ia
0-57
0-27
0- 81
0-52
4
0-80
0-72
0-38
0-46
0°25
0-46
0-51
5
0-74
1°2
0»56
0-77
0-7
0-7
0-761
6
0-79
0-62
0-63
0-22
0-85
0-518
7
0-71
0-60
0-47
1-1"
1-0
0-78
0-777
Average
Assessment .
of each Receiver
0'71
0-623
0°272
0-704
0-45
0-70
0- 576
for All Observers
Each figure shown under the six columns marked "Receivers'' is the ratio of the
distance at which the dot structure was just visible to the distance at which the
raster lines were just visible.
3. RESULTS AND CONCLUSIONS.
Table 1 shows the results of the objective tests on the sound channels of the
receivers .
Ail recorded figures are expressed in decibels below the listeners 7 prefer-
ence . sound level used for aligning the receivers. (See Research Department Report
No. A. 026. }
The columns were recorded with unmodulated sound carrier present. The
column headed "With Picture" may be interpreted as a measure of the normal breakthrough
of vision on sound for the receiver tested, and it will be noticed that, whereas
Receiver No. 1 showed negligible breakthrough, Receiver No . 5, the rejectors of which
were maladjusted, suffered consider arly from this defect,
when the chrominance subcarrier was substituted for picture modulation, even
less breakthrough than that from the normal picture signal resulted. The relevant
figures are given in the final column. This may be interpreted as an indication of
the response of most domestic receivers, which is usually attenuated considerably at
frequencies beyond about S§ Mc/s.
From Table 2, recording the visual tests, it can be seen that the dot
structure was less visible than the normal raster lines present on the screens.
The figure in the lower right hand corner of the table may be termed "The
ratio of dot visibility to raster line visibility for an average observer on an average
receiver" within the limits of the data available.
This figure of 0*576 indicates that an average observer, situated at such a
distance that the identity of the raster lines was just visible, would need to advance
to approximately half this distance before being able to appreciate the dot structure
due to chrominance subcarrier interference.
APPENDIX 1
INCREASE IN APPARENT BRIGHTNESS ;"■:? CHROMINANCE SUBCARRIER DOT STRUCTURE
DUE TO NON-LINEARITY OF PICTURE TUBE TRANSFER CHARACTERISTIC
Recent tests have revealed that the brightness of those portions of the
screen of a picture tube in which chrominance subcarrier is present is greater than
that of a comparison patch having the same luminance but no chrominance contribution.
Whilst this may be self— evident to some observers, it is possible to be
misled by the following argument:
Suppose a colour television signal S of N.T«S»C. type is being received
having a luminance a and a chrominance am, where m is the chrominance— to- luminance
amplitude ratio. We may write for the signal on the control electrode of the receiver
picture tube
S = a(l + m sin x) ( 1}
where x - 27fft, f being the frequency of the chrominance subcarrier.
The brightness on the tube screen will be proportional to S 2 * 5 if 2 3 5 is the
exponent of the tube transfer characteristic assumed to be of simple power-law form.
If, as we shall assume later, the transfer characteristic of the eye in terms of
sensation units of brightness as a function of objective brightness, iu, say, ft—
lamberts, is also a simple power law* of exponent approximately 1/2*5, the sen sat. ion
of brightness might be thought to be proportional to
(32- 6) l/2- 5 = g
and therefore no change in signal brightness would occur as a result of tube transfer
characteristic curvature.
The above argument, however, represents the case in which the viewer 1 s eye
is so close to the tube screen that the variations in brightness due to the term
am sin x in equation 1 are clearly visible so that the mean value may be estimated.
At normal viewing distances this is by no means true; the viewer*s eye integrates
over an area containing many dots. We observe an apparent brightness which is the
2 s 5 th root of the mean value of S 2 * 5 * We have to find a mean brightness sensation
given by
r &r U/2«5
B ®i^= I a*. 5 (1 + m sin x)>2* 5 ^ I (2)
In the absence of dot structure the brightness sensation given by the luminance a is
simply B ^ a. The ratio of brightness sensation with dot structure to that without
is therefore
~ 27T . "~
■i [ (1 + m sin xi 2,5 dx
_
R = B /B =
1/2* 5
(3)
In the case of a saturated green signal we have, very roughly, m = 1, and
1-3.
During reception of test transmissions of a synthesised saturated green
signal from the Alexandra Palace television transmitter, the brightnesses of large
areas of the screens of receiver picture tubes were measured and compared with those
corresponding with the luminance component only. Objective brightness ratios of the
order of 1*5 to 1 were obtained from the various receivers used. The subjective
equivalent of this ratio is 1*18 to 1, in sensation units. Prom separate tests it
was known that the television transmitter attenuated the chrominance sub carrier by
about 3 dB and raised the luminance component of the composite signal by about li dB
with respect to a separate luminance component transmitted solely for comparison
purposes. We thus have for this case m = 0*71 (-3 dB). R = 1* 15 from equation 3,
but we must add the l£ dB of "lift" inserted by the transmitter and mentioned above,
thus we obtain a calculated value for R of 1*37.
It would therefore appear that this explanation for the increased brightness
of areas covered by the dot structure due to chrominance subcarrier is only qualita-
tively true, since measurements yielded sensation ratios of the order of 1*18 and
calculation gave a figure of 1*37.
*B, B* C. He sear ch Department Report Mo. T.044,Fig.2.
It is interesting to note, however ? that those receivers which attenuate
signals at the higher video frequencies, such ?.s t !:-.e chrominance 'subcarrier, show a
ratio between dot structure brightness .and comparative luminance— patch brightness of
near unity; though the better the frequency response of the receiver the greater
will be the ratio R, since the term am sin x applied to the non— linear picture— tube
characteristic will be larger* This probe... I y accounts for the fact that measured
ratios were less than calculated values. In the case of colour television receivers
fitted with subcarrier attenuation filters, the dot structure will be absent and so
coloured portions of the pictures will have their true brightnesses, the ratio R
being unity*
It is thus seen that the presence of the dot structure on the screen of the
picture tube, whether it be monochrome or polychrome, introduces a panchromatic error
and destroys the constant luminance feature* (when it exists) of the N.T.S.C. colour
television system.
Another interesting corollary of the rectification of subcarrier by picture-
tube non-linearity is that electrical loss of resolution (attenuation of the higher
video frequencies) can be distinguished from optical loss of resolution. If the dot
structure is not itself visible but the patch on the tube screen where it should
appear is too bright, that is, brighter than its true luminance value, then subcarrier
must be present in the electron beam scanning the phosphor ^screen, but as it is not
actually visible upon close inspection, optical loss must be responsible for its
non— appe ar ance ,
*For another cause of failure of constant luminance see B.B.C. Research Dep artment Technical
Memorandum No. T, 10 12,
APPENDIX 2
THE PERCEPTIBILITY OF DISCRETE RASTER LINES UNDER CONDITIONS
OF LOW .AMBIENT LIGHTING
Screen brightness, apostilbs
20 30 40 60 80 I0O
It
(0
9
1
1
l
1
i
1
1
1
1
-
e
_
7
6
5
-
-3-5
-3-0
2-5 t
w
5
2 3 4 6 6 10
Screen brightness, toot lomberts
This test was performed in order
to determine the distance at which raster
lines are still just discernible to a
critical viewer. The brightness of the
raster was varied in discrete steps and
measured by a photometer. At each level
of brightness, observation was. made of the
distance at which the raster lines were
just visible. During the experiment, the
ambient lighting was maintained at a low
and constant level.
The curve shown is the average of
four series of observations, up^^' 16 in.
screen,